\n– A human<\/strong><\/em><\/p>\n
Now imagine Google Page Speed Insights is that tool. Imagine it\u2019s a speedometer to see who\u2019s fastest, who crosses the finish line first.<\/p>\n
\u201cI\u2019ve seen that my competitor\u2019s website is better than mine!! Google says so!!\u201d<\/strong><\/em> If you work in web layout and\/or online marketing, this will sound very familiar. You\u2019re not imagining it \ud83d\ude09 It\u2019s happened to me too, which is why I decided to dig a little deeper to understand the tool and see what we can do to improve our results with it. NOTE:<\/strong> This post is based on true events. Everything you\u2019re about to see really happened, in exactly the order it\u2019s written.<\/em><\/p>\n <\/p>\n The whole analysis is based on comparing a real website with a clone of it, hosted on the same server under the same conditions: same data as the real site, same code, same active plugins, etc. In theory they\u2019re the same website. NOTE: the result on the left is the real website<\/strong>, the one on the right is the cloned website<\/strong> I\u2019ll be experimenting on. This will be the case throughout the post.<\/em><\/p>\n Since the tool runs a fresh analysis and calculates on the fly each time, I run them again to make sure we\u2019re on a level playing field, more or less\u2026 and this is the result I get:<\/p>\n As you can see, the results are very similar in both cases.<\/p>\n NOTE:<\/strong> if you run it again immediately, it returns the last cached analysis, so make sure to switch away from the page and wait a while or analyze another URL<\/em><\/p>\n <\/p>\n From this last test I also extract the main data for the experiment, which look like this:<\/p>\n <\/p>\n \nSo as not to stop at the homepage, I also analyze a page with little content and mostly plain text, such as the legal notice.<\/p>\n With no reasonable explanation, this is the result:<\/p>\n Yet the rest of the analysis data are very similar:<\/p>\n <\/p>\n \nI run the analysis on both pages again, and this is the result:<\/p>\n <\/p>\n First conclusion:<\/strong> <\/p>\n The website I\u2019m using for the analysis is the very opposite of minimalist. It isn\u2019t clean, it doesn\u2019t breathe… and if there were a gap anywhere, we\u2019d still see if we could fill it with another banner\u2026 and if things move a bit, even better\u2026<\/p>\n It reminds me a bit of this: Currently, besides its own requests and those to our CDNs, this website makes requests to 18 external domains to collect data and\/or use third-party tools for chat, metrics, smart search and \u201cmarketing automation,\u201d among others\u2026<\/p>\n One of our first suspicions when we set out on this analysis was that the tools built into the website could interfere with the test result depending on how they were implemented and\/or by their very nature.<\/p>\n Now I\u2019ll gradually disconnect tools and JavaScript tracking codes on the cloned website and again compare both sites at the same moment.<\/p>\n <\/p>\n First, I choose to disable this tool, which works as a JavaScript plugin on the store. I check that the cloned site no longer sends requests to the search tool\u2019s domain and then analyze both sites at the same time. Since I can\u2019t believe my eyes\u2026 I run it again:<\/p>\n (Logbook. Day 1. Note 1. I\u2019m going crazy. I don\u2019t know if I can take this one more minute\u2026)<\/em><\/span><\/p>\n Since I\u2019m blown away, I run it again:<\/p>\n (Logbook. Day 1. Note 2. This job is definitely too hard for me\u2026)<\/em><\/span><\/p>\n <\/p>\n Second conclusion:<\/strong> <\/p>\n And this analysis could end here… I don\u2019t know if it\u2019s worth wasting more time on something that isn\u2019t reliable\u2026 Still, I give it one last chance and try again\u2026 and this is the result:<\/p>\n (Logbook. Day 1. Note 3. I really don\u2019t know why I keep going with this\u2026 I\u2019m seriously rethinking my future\u2026)<\/em><\/span><\/p>\n Oh come on, one last time and I\u2019ll stop\u2026 final result:<\/p>\n So I don\u2019t keep banging my head against the wall, I\u2019ll do what Clemente\u2019s teams did, \u201ckick it long and move on,\u201d and assume the change hasn\u2019t improved the cloned site\u2019s results compared with the original.<\/p>\n <\/p>\n Next I disable the plugin that lets shoppers leave reviews about the website and their purchases. I check that the cloned site no longer sends requests to that domain, and then analyze both sites at the same time.<\/p>\n This is the result:<\/p>\n Unbelievable… I remove things from the cloned site and the real one \u201cimproves\u201d… I had to double-check I wasn\u2019t disconnecting things in the wrong place… but no…<\/p>\n (Logbook. Day 1. Note 4. I should have taken the blue pill, I\u2019d be much happier in a fake reality\u2026)<\/em><\/span><\/p>\n I try again, and this is the result:<\/p>\n And one last time:<\/p>\n And so I don\u2019t go any crazier\u2026 I\u2019ll carry on as if I hadn\u2019t seen a thing\u2026<\/p>\n <\/p>\n The next disconnection concerns metrics for social media traffic and conversions. I check that the cloned site no longer sends requests to the social network\u2019s domain, and then analyze both sites at the same time.<\/p>\n This is the result:<\/p>\n And right now I feel like this:<\/p>\n I run both tests again:<\/p>\n (Logbook. Day 1. Note 5. If I don\u2019t make it out alive, tell my family I love them\u2026)<\/span><\/em><\/p>\n At this point in the experiment, the cloned site has slightly reduced its requests and page weight. <\/p>\n Now I\u2019m going to disconnect a general metrics tool (not Google Analytics). I check that the cloned site no longer sends requests to the tool\u2019s domain, and then analyze both sites at the same time.<\/p>\n This is the result:<\/p>\n It must be rush hour at AWS right now…<\/p>\n I\u2019m going to try again… I don\u2019t even know why… but I do:<\/p>\n Ignore it\u2026 moving on\u2026<\/p>\n <\/p>\n Now I\u2019ll disconnect the first marketing automation tool, one of those that do everything on your website so you automatically sell more and get more customers and more of everything\u2026 Promise! This is the result:<\/p>\n And, still not knowing why, I run it again:<\/p>\n (Logbook. Day 1. Note 6. I hear nothing, there\u2019s only silence\u2026 I see nothing, there\u2019s only darkness\u2026)<\/span><\/em><\/p>\n <\/p>\n Now I\u2019ll disconnect the second marketing automation tool, which is also one of those that do everything on your website so you automatically sell more and get more customers and more of everything\u2026 but better\u2026 or not\u2026 What do I know? This is the result:<\/p>\n It must be a mirage\u2026 an oasis, it has to be\u2026 I try again:<\/p>\n Money can\u2019t pay for this\u2026 I\u2019ve seen gentler torture\u2026 there was no need for this\u2026<\/p>\n At this point, the cloned site has dropped about 40 requests and a couple of MB in weight compared with the original (although the transferred size hasn\u2019t dropped as much)<\/p>\n <\/p>\n I check that the cloned site no longer sends requests to this tool, and then analyze both sites at the same time.<\/p>\n This is the result:<\/p>\n More of the same\u2026 I\u2019ve been disconnecting things for a while, I swear… but I\u2019m right where I started\u2026 I\u2019m beginning to suspect I\u2019m looking in the wrong place\u2026<\/p>\n I run it again:<\/p>\n Do I believe it? Or don\u2019t I? My gut says no, the evidence says no, life says no\u2026 so why believe it?<\/p>\n (Logbook. Day 1. Note 7. I\u2019m seriously considering death as the solution to all this\u2026 A strange force inside me pushes me to run the test again:<\/p>\n I refer you to the first two conclusions for my opinion on what I\u2019m witnessing.<\/p>\n <\/p>\n This website uses a well-known chat plugin to offer this feature and serve its customers. I disconnect the tool, check that the cloned site no longer sends requests, and then analyze both sites at the same time.<\/p>\n This is the result:<\/p>\n Holy crap!<\/strong> (pardon my French) I\u2019ve broken the psychological 40 barrier \ud83d\ude42 Just like with age, that\u2019s where life really begins… but… should I believe my eyes? Or is it still a Google Skynet hoax to distract me?<\/p>\n God (I mean\u2026 Google), if you\u2019re messing with me, you should know it\u2019s not funny at all. (Logbook. Day 1. Note 8. Mental note: start using Yahoo as my search engine)<\/em><\/span><\/p>\n Once again, I run it, just in case. It\u2019s not that I don\u2019t trust it… you know\u2026<\/p>\n And\u2026 one last time:<\/p>\n I know I shouldn\u2019t get my hopes up\u2026 but I like what I see… who could resist? After all, sight is the most influential of the senses… I\u2019ll clean my glasses, just in case…<\/p>\n Out of curiosity, I run it once more:<\/p>\n (Logbook. Day 1. Note 9. What\u2019s that light? It\u2019s drawing me in… I can\u2019t resist\u2026)<\/em><\/span><\/p>\n <\/p>\n At this stage of the test, only Google-related trackers remain on the site. Page weight and transferred data have dropped considerably, and the site loads much lighter.<\/p>\n Although I don\u2019t believe any of it, since the tool has proven very erratic, disconnecting the chat tool seems to have brought some stability to the results, with the cloned site improving over the real one (and over its own initial results)<\/p>\n NOTE:<\/strong> I should say that at this point the test takes about half as long to run on the cloned site. (Which makes sense, I suppose\u2026)<\/p>\n <\/p>\n Inside GTM we have Analytics, Google Ads and the other related tools.<\/p>\n I disconnect the tracker, check that the cloned site no longer sends requests to any Google domain, and then analyze both sites at the same time.<\/p>\n This is the result:<\/p>\n It looks consistent, within its random chaos\u2026<\/p>\n <\/p>\n And since I\u2019m starting to believe what I see, despite past bad experiences, I want to confirm it in another scenario. So I\u2019ll retest the low-content pages I analyzed at the start, but in this new setup:<\/p>\n And this is where I\u2019m once again amazed by the tool\u2019s scientific rigor\u2026<\/p>\n Umm… does anyone have an explanation? Why so much pain?<\/p>\n What have I done to the world? Why me?<\/p>\n Despite the real site\u2019s result, this is closer to what I expected to find…<\/p>\n It looks consistent… although once again the real site has dropped well below what was expected… though with this tool, I no longer know what to expect…<\/p>\n I\u2019ll test the homepage again and see what I get:<\/p>\n It\u2019s just as I left it… As if nothing had happened\u2026 As if the tool knew what it was doing\u2026<\/p>\n <\/p>\n If we trust the results, the clearest conclusion would be that the chat is the third party that most affects the result when third parties are disabled.<\/p>\n I\u2019ll try to confirm whether this conclusion is correct. To do so, I\u2019ll reconnect everything I disconnected except the chat tool. We\u2019re back to figures similar to the start in terms of weight and requests.<\/p>\n And I test it:<\/p>\n WTF? Seriously\u2026 this isn\u2019t normal\u2026 let\u2019s see, again:<\/p>\n Now this looks more like what I expected\u2026 But who can trust this?<\/p>\n Well\u2026 one last time…<\/p>\n <\/p>\n I reconnect the chat tool to return to the initial setup.<\/p>\n And I run the tests:<\/p>\n (Logbook. Day 1. Note 10. The hallucinations are back\u2026)<\/em><\/span><\/p>\n The real store\u2019s result is absurd\u2026 so I run the test again\u2026<\/p>\n I can\u2019t use the word \u201cnormal,\u201d but this is closer to what I expected\u2026<\/p>\n Well… it seems clear then… even so, I disconnect it again from the cloned site and test:<\/p>\n Suspicious… The expected result on the first try! Let\u2019s see… again:<\/p>\n Why, chat tool?! Why?!! You were supposed to be here to help me!! So, after my dramatic performance, here\u2019s the third conclusion of this study:<\/p>\n <\/p>\n Third conclusion:<\/strong> <\/p>\n Okay, that chat tool is the devil, and that\u2019s been proven. But are we going to settle for that?<\/p>\n Clearly, apart from the chat, the other third parties interfere with the result to an extent that can be disregarded in an optimization study.<\/p>\n So, on to the next experiment.<\/p>\n <\/p>\n The first thing I ask myself is: how far can we go? 99? 100?<\/p>\n This first experiment involves leaving the store code \u201cempty,\u201d so it loads all the common JS and CSS our system needs to run, along with the store\u2019s own code, but renders nothing.<\/p>\n The result is a blank page with all the trackers and the common CSS+JS base. At this point, the chat tool is the only one that isn\u2019t active.<\/p>\n And the test result is surprising:<\/strong><\/p>\n I run it a couple more times, since it\u2019s already been proven that averaging the results is the way to go…<\/p>\n I wonder what percentage impact the trackers still connected have now. So I set out to disconnect them all again. ALL at once.<\/p>\n Weight and load times are now at a minimum\u2026 And the result is still surprising\u2026 it\u2019s not a 100!<\/p>\n <\/p>\n So I decide to take it one step further… I remove the store\u2019s cookie policy and GDPR setup (which displays content dynamically with JavaScript) and put in some plain HTML \u201ccontent.\u201d<\/p>\n The result is a website that neither reads nor creates cookies and renders plain text directly from the HTML instead of via JS.<\/p>\n I want to believe this is the minimum we can get from a \u201cstore\u201d built with our system (even if it isn\u2019t realistic)<\/p>\n It would be the equivalent of having an HTML file on the server and reading it, but going through our server language, creating a user session and all the other necessary processes.<\/p>\n (Logbook. Day 1. Note 11. Twice nothing is nothing\u2026 that\u2019s what Google has taught me)<\/em><\/span><\/p>\n But is this good for anything? Well, we\u2019re closer to something… whatever it is\u2026<\/p>\n <\/p>\n To rule out web server and\/or database processing times as a factor, I create an HTML file with the same content as the dynamically rendered page and run a test on it.<\/p>\n The result is almost identical, except that in the screenshots we already have rendered \u201ccontent\u201d from the very first one.<\/p>\n Which makes me wonder:<\/strong> why does it say it takes 1.7s to deliver the first contentful paint if it\u2019s plain HTML?<\/p>\n So I remove the JS loading and try again.<\/p>\n Okay… 100… But the tool took 1.6s to render??<\/p>\n And without CSS??<\/p>\n Wow, that\u2019s fast… but personally, 0.6 seconds to render a practically empty HTML page seems like a lot… I take it this is the \u201cminimum\u201d time we can aim for with this tool…<\/p>\n And while we\u2019re at it\u2026 with JS but without CSS?<\/p>\n It takes the same time for the first render, but 2.6s to become interactive. <\/p>\n Okay, I\u2019ve reached the lowest common denominator. Now I need the greatest common multiple.<\/p>\n (Logbook. Day 1. Note 12. Okay… I\u2019m delirious again… I should stop before it\u2019s too late\u2026)<\/em><\/span><\/p>\n I\u2019ll restore the homepage as it was and gradually remove elements to see what impact they have. Remember that at this point the site has no trackers or JavaScript tools built in.<\/p>\n The only thing I can\u2019t remove is our common JS and CSS code. In the previous experiments we saw the minimum to expect in this case:<\/p>\n <\/p>\n I remove all the JavaScript that comes from the store itself. That way the site is complete, but only has the JS from our common system engine.<\/p>\n And this is the result:<\/p>\n <\/p>\n I remove all the CSS and also the JavaScript coming from the store. That way the site is complete, but only has the JS\/CSS from our common system engine.<\/p>\n And this is the result:<\/p>\n Not bad at all either\u2026 but… something doesn\u2019t add up\u2026<\/p>\n <\/p>\n At this point I remember we have a demo base code that we build stores and experiments on. Clean but complete, functional code.<\/p>\n I set up a new environment with a store running that code so I can test it with this tool. And\u2026<\/p>\n This site also loads CSS and JS from our common engine. It also has its own CSS and JS, also runs actions when the DOM is \u201cready\u201d and has dynamic elements\u2026 So? What\u2019s going on?<\/p>\n <\/p>\n Having made it this far still sane, which is quite an achievement, all that\u2019s left is to put the icing on the cake with a grand finale: dramatic, exciting, with a touch of suspense and, above all, \u201crevealing.\u201d<\/p>\n What if the culprit is\u2026 the design!? Well\u2026 responsive design! Well… not exactly… or yes… Not the concept, but how it\u2019s \u201cunderstood\u201d in many cases and where clients push us…<\/p>\n (Admittedly, Google\u2019s tool has a curious way of measuring, which I\u2019ll try to explain later with some conspiracy theories)<\/p>\n And to try to prove it, the best thing is to run another experiment, like this one:<\/p>\n The website I\u2019m analyzing has a complex mega menu. So complex that its \u201cmobile conversion\u201d isn\u2019t possible because the hierarchies are very different. (It takes up about 700 lines of HTML and loads 54 images)<\/p>\n So the site actually has 2 menus, and one or the other is shown depending on the device. This means that on mobile, all the HTML, images and everything else from the desktop menu are loaded even though they aren\u2019t used.<\/p>\n What happens if I remove the desktop menu from the code? At this point, the code contains all the JS and CSS, except for the plugins and the chat tool.<\/p>\n Suddenly the mobile rendering engine seems lighter, fewer images and less HTML are loaded, and the result is better.<\/p>\n I got a better result than without loading any JS on the site\u2026<\/p>\n (Logbook. Day 1. Note 13. I can no longer tell good from bad, so everything is good)<\/em><\/span><\/p>\n <\/p>\n From here, all I can do is go back to building an HTML file from the dynamically generated result, to try out changes that might optimize the result so far.<\/p>\n I start from the previous experiment (site with no chat, no trackers and no desktop menu)<\/p>\n First of all, I check that the result is still as expected:<\/p>\n Move all the CSS to the very end of the code:<\/p>\n The result improves, but at the cost of the site spending a bit longer unstyled and looking \u201ca little ugly.\u201d Make the CSS calls asynchronous<\/p>\n The result improves slightly, but at the cost of the site spending a bit longer unstyled and looking \u201ca little ugly.\u201d<\/p>\n Make the JS calls asynchronous<\/p>\n Once again, the result improves slightly, but at the cost of the site spending a bit longer unstyled and looking \u201cugly.\u201d<\/p>\n And what if I add the chat back on top of all this?? I\u2019ll try it a couple of times and see\u2026<\/p>\n Lucifer!!!! Get away from my website!!!!<\/strong><\/p>\n <\/p>\n During this odyssey through the cyberspace of load speed, we may have lost touch with all reality. That\u2019s normal. It\u2019s what Google would want.<\/p>\n (Logbook. Day 2. Note 1. I didn\u2019t come this far for nothing\u2026 future generations will value all this\u2026)<\/em><\/span><\/p>\n With everything we\u2019ve \u201clearned\u201d so far, which is quite a lot, we just need to get back to reality. What can we do to improve the score of the site we\u2019re analyzing?<\/p>\n Once again, all I can do is go back to building an HTML file from the real site\u2019s output, to try out changes that might optimize the result.<\/p>\n This time I start from scratch. That is, the site has all its trackers and plugins, chat included.<\/p>\n First of all, I check that the result is still as expected:<\/p>\n So far so good, no \u201csurprises.\u201d The tool is doing what it\u2019s supposed to, which is a welcome change at this point\u2026<\/p>\n I\u2019ll put the CSS at the very end and see what I get. I try it a couple of times:<\/p>\n What\u2019s going on? Was yesterday April Fools\u2019 Day or something? Suddenly the tool has decided to be \u201cstable\u201d\u2026<\/p>\n Obviously, the aesthetic \u201cproblem\u201d of the site not being properly rendered for a while appears, which is \u201cugly\u201d but could be worked around. Now I\u2019ll make the CSS calls asynchronous. It shouldn\u2019t matter much once the CSS is at the end of the load, but since it supposedly \u201cscores,\u201d we\u2019ll do it\u2026 \ud83d\ude42<\/p>\n Let\u2019s draw a stupid veil over that\u2026<\/p>\n Now I\u2019ll make the JS calls asynchronous.<\/p>\n Let\u2019s draw another stupid veil over that\u2026<\/p>\n (Logbook. Day 2. Note 2. Every fall makes me stronger\u2026 every mistake, wiser\u2026 every attempt, more unhinged\u2026)<\/em><\/span><\/p>\n What if I remove the chat?<\/p>\n What if I remove the desktop mega menu?<\/p>\n WTF?? Have I mentioned I hate this tool? Google\u2026 I curse you!<\/p>\n <\/p>\n Fourth conclusion:<\/strong> <\/p>\n Okay, I\u2019m going back\u2026 I\u2019ll leave the mega menu and everything else as it was\u2026 all of it\u2026<\/p>\n Now I\u2019ll do what Google would really like this page to do\u2026<\/p>\n Out with the CSS! Out with the JS!<\/p>\n Welcome to the Age of AMP Ultron<\/p>\n None of it matters: image weight, number of images, HTML weight\u2026 Nothing!! This site weighs almost 2MB and makes 111 requests!!<\/p>\n The only thing that matters is how long the rendering engine takes to make the site available to the simulated user.<\/p>\n It was right in front of us, and we didn\u2019t want to see it\u2026<\/p>\n (Logbook. Day 2. Note 3. They\u2019ve been playing with me… They knew everything from the start!)<\/em><\/span><\/p>\n <\/p>\n This whole psychedelic trip couldn\u2019t be for nothing. It was essential to learn something\u2026 to gain knowledge and be able to apply it. And, above all, to tell the fine line between madness and sanity.<\/p>\n <\/p>\n Simplicity, and for users to have the content available as soon as possible and be able to interact with it. The more complex a page is, the lower its score, unless you try to \u201cdisguise\u201d it from the measuring tool.<\/p>\n By sacrificing initial interactivity and site elements to simplify it, and accepting that it may not look right at first. By deferring loads until after the first \u201cuser-interactive\u201d result.<\/p>\n A dedicated mobile version would help produce cleaner code for this test and get a better score. Especially if, by design, the site\u2019s responsive layout isn\u2019t \u201cvery natural\u201d and many elements are duplicated because they can\u2019t be \u201ctransformed\u201d dynamically.<\/p>\n Making that mobile version AMP would help even more.<\/p>\n As with Google and SEO rankings, this tool has dozens of variables to analyze. It isn\u2019t all JS and CSS, but those are what \u201cslow things down\u201d the most.<\/p>\n Simplifying the site helps, but if you upload two 2MB images, you\u2019ll be penalized\u2026 so not everything that can be improved comes down to code.<\/p>\n Proof of this is that the \u201cbase\u201d code scores very well, yet it carries all the same base JS\/CSS code as the store we analyzed\u2026<\/p>\n Why? Because it\u2019s very simple. As soon as we start complicating a site\u2019s responsive design \u2014mega menu, sticky header, interactive search, dozens of banners + categories + products on the homepage, JS-generated dynamic elements such as popups, cookie consent, etc.\u2014 we introduce elements that slow down the tool\u2019s rendering engine.<\/p>\n In my view, and this is just a theory, the tool\u2019s job is to give you a score based on the \u201cimprovement\u201d it believes can be made in code and resources compared with the result it obtained. Something like: \u201cIf I had to build a site like the one I\u2019m looking at, I\u2019d do it this way\u2026 and you did it that way\u2026 I compare them and tell you the difference\u2026\u201d.<\/p>\n That\u2019s why a horribly heavy, ugly site with no CSS\/JS to interpret gets a score of 100. Google couldn\u2019t do it any better! \ud83d\ude00 (Probably not any worse either\u2026)<\/p>\n At this point, I do what I should have done at the start: read the instruction manual. Here\u2019s an excerpt from Google\u2019s official documentation on this tool:<\/p>\n https:\/\/developers.google.com\/speed\/docs\/insights\/about?hl=es-419 <\/a><\/p>\n \u201cPSI checks to what extent a page follows common performance best practices and scores it from 0 to 100, a rating that roughly indicates its performance headroom\u201d<\/p>\n And further: \u201cThe optimization score is a rough estimate of the performance headroom of a page\u201d<\/p>\n And to finally prove that this tool doesn\u2019t measure our site\u2019s speed and isn\u2019t an indicator of whether other sites are faster than ours:<\/p>\n \u201cA slow page may have a high optimization score because the ratio of render-blocking resources to the page\u2019s total resources is low. The opposite can also happen: a fast page may have a low optimization score because it has many render-blocking resources relative to its total resources\u201d<\/p>\n Phew\u2026 if I\u2019d read the manual before starting, I wouldn\u2019t have gone through this ordeal\u2026<\/p>\n But why are the results so different for the same test?<\/p>\n My theory is that, depending on when you run the test, the server running it takes more or less time to execute it and, in some cases, to reach all the domains it has to request, and that ends up blocking load time… But it\u2019s just a theory…<\/p>\n Another could be that the test actually does nothing and spits out a random number based on total load time… That\u2019s all… xDD And that behind it all there\u2019s a toy monkey clapping cymbals\u2026<\/p>\n Yes and no.<\/strong><\/p>\n What did you expect? After the madness I\u2019ve been through\u2026<\/p>\n It\u2019s important to at least take a look and check we haven\u2019t made mistakes that penalize us needlessly. Very heavy images, techniques that could be improved\u2026<\/p>\n But at the same time, it\u2019s important not to obsess over the results. After all, it isn\u2019t a stable tool. And its result depends on variables that are often outside your control.<\/p>\n So you might ask: is it worth investing to get a better score?<\/p>\n And once again my answer is: it depends. To fix a disaster, obviously yes. It\u2019s a sign something isn\u2019t quite right. To beat others: obviously not. In that case you\u2019re not comparing like with like.<\/p>\n And finally, let\u2019s not forget: improving your homepage score won\u2019t change the score for the rest of your site. Google is experimenting with using this tool across your site\u2019s navigation, ALL of it, not just the homepage.<\/p>\n So if you want to focus on something, aim for structural, site-wide improvements that genuinely optimize your website, not just to stay \u201cahead of others.\u201d<\/p>\n Now that I\u2019ve drawn my own conclusions, I looked around to see if anyone else had reached the same ones. Here\u2019s some further reading, if you haven\u2019t had enough:<\/p>\n They\u2019re really interesting, trust me.<\/p>\n (Logbook. Day 2. Note 4. All things come to an end\u2026 I hope this isn\u2019t mine)<\/em><\/span><\/p>\n I hope you\u2019ve enjoyed reading this post as much as I enjoyed writing it. Enjoying something so tedious must be worthy of study.<\/p>\n","protected":false},"excerpt":{"rendered":" Imagine we have a tool we believe is useful and will help us improve sales. Imagine, too, for a moment, that we can use that tool on a competitor\u2019s website and \u201ccompare ourselves\u201d to see who\u2019s better. Who wouldn\u2019t want something like that? But this completely imaginary situation comes with something very real, very human, […]<\/p>\n","protected":false},"author":2,"featured_media":14576,"comment_status":"open","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[2519],"tags":[],"class_list":["post-18995","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-technology"],"yoast_head":"\n
\n– A human with a website<\/strong><\/em><\/p>\n
\nSince I don\u2019t know how the tool works internally or what criteria it uses for testing, I\u2019m going to dive right in with a series of tests to figure out what it expects from a website. So as not to bias my opinion, I\u2019ve decided not to look anything up online or read the official documentation (instruction manual? Nah!) and to do this testing \u201cblind.\u201d I think it\u2019s more fun that way.<\/p>\nTest \/ comparison conditions (first conclusion)<\/h2>\n
\nWith this setup, I run the first Google Page Speed analysis on both sites and get the following mobile result:<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/02.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/03.png\")
<\/p>\nREAL WEBSITE<\/strong>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/03-1.png\")
<\/pre>\nC<\/strong>LONED WEBSITE<\/strong>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/04.png\")
<\/pre>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/005.png\")
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/006.png\")
<\/p>\nREAL WEBSITE<\/strong>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/007.png\")
<\/pre>\nC<\/strong>LONED WEBSITE<\/strong>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/008.png\")
<\/pre>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/009.png\")
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/010.png\")
<\/p>\n
\nThis tool is less stable than an internet connection in China<\/p><\/blockquote>\nThe case at hand<\/h2>\n
\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/tenor.gif\")
<\/p>\nSmart search<\/strong><\/h4>\n
\nThis is the result:<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/011.png\")
<\/p>\n<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/012.png\")
<\/p>\n
\nThis tool is as credible as a politician on the campaign trail<\/p><\/blockquote>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/013.png\")
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/014.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/015.png\")
<\/p>\nRatings and reviews<\/strong><\/h4>\n
<\/p>\n<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/016.png\")
<\/p>\nSocial media metrics<\/h4>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/017.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/g9wdbemjy8ks0mz6ltvn.gif\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/018.png\")
<\/p>\n
\nOnward! (What choice do I have\u2026)<\/p>\nGeneral metrics<\/h4>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/019.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/020.png\")
<\/p>\nMarketing automation 1<\/h4>\n
\nI check that the cloned site no longer sends requests to the tool\u2019s domain, and then analyze both sites at the same time.<\/p>\n<\/p>\n<\/p>\nMarketing automation 2<\/h4>\n
\nI check that the cloned site no longer sends requests, and then analyze both sites at the same time.<\/p>\n<\/p>\n<\/p>\nMore general metrics<\/h4>\n
<\/p>\n<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/021.png\")
<\/p>\n
\nIn the event of a murder trial, this should be considered a mitigating circumstance.)<\/em><\/span><\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/022.png\")
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/023.png\")
<\/p>\nChat tool<\/h4>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/024.png\")
<\/p>\n<\/p>\n
\nMy sanity is at stake, and this isn\u2019t helping.<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/025.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/026.png\")
<\/p>\n<\/p>\nCurrent situation after disconnecting third parties en masse<\/h2>\n
Google Tag Manager<\/h4>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/027.png\")
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/028.png\")
<\/p>\n<\/p>\nA comparison from another angle<\/h2>\n
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/029.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/030.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/031.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/032.png\")
<\/p>\nReconnecting everything<\/h2>\n
<\/p>\n<\/p>\n<\/p>\n<\/p>\nWhat if we go back to the start? (third conclusion)<\/h2>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/033.png\")
<\/p>\n<\/p>\n<\/p>\n<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/034.png\")
<\/p>\n
\nI can\u2019t believe it, I give up\u2026 I have to try again, I can\u2019t live without you!!<\/p>\n
\nUsing certain chat tools on your website is like driving a semi truck to work:
\nYou\u2019ll get there\u2026 but\u2026 at what cost? And\u2026 where will you park?<\/p><\/blockquote>\nWhat else can we do?<\/h2>\n
The best we can hope for? First experiment<\/h2>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/035.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/036.png\")
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/037.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/038.png\")
<\/p>\nSecond experiment<\/h2>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/039.png\")
<\/p>\nThird experiment<\/h2>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/040.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/041.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/042.png\")
<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/043.png\")
<\/p>\n
\nWith this I have enough information to move on to more \u201crealistic\u201d tests.<\/p>\nNow the same, but the other way around<\/h2>\n
\n
Fourth experiment<\/h2>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/044.png\")
<\/p>\nFifth experiment<\/h2>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/045.png\")
<\/p>\nComparisons are odious<\/h2>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/046.png\")
<\/p>\nThe harsh reality (sixth experiment)<\/h2>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/047.png\")
<\/p>\nThe ultimate experiment?<\/h2>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/048.png\")
<\/p>\nPart 1<\/h4>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/049.png\")
<\/p>\n
\nThe first render is fast; everything else stays as before.<\/p>\nPart 2<\/h4>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/050.png\")
<\/p>\nPart 3<\/h4>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/051.png\")
<\/p>\nPart 4<\/h4>\n
<\/p>\nBack to reality (fourth conclusion)<\/h2>\n
<\/p>\n<\/p>\n![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/052.png\")
<\/p>\n<\/p>\n<\/p>\n<\/p>\n<\/p>\n
\nThis tool is like a kid with money in a candy store: it knows what it wants, but it doesn\u2019t know what it wants\u2026<\/p><\/blockquote>\nPart 8 \u2013 What Google would like\u2026<\/h2>\n
![\"\"](\"https:\/\/trilogi.com\/blog\/wp-content\/uploads\/sites\/2\/053.png\")
<\/p>\nTa-daaaaaaaa!!!!<\/strong><\/h4>\n
From madness to sanity: the commandments<\/h2>\n
What does this tool expect from our pages?<\/h4>\n
How do we achieve that?<\/h4>\n
Are there other alternatives?<\/h4>\n
How on earth does this tool score?<\/h4>\n
Should I pay attention to this tool?<\/h4>\n
A bit more reading to wrap up<\/h4>\n
\n